Automated
Customer Self Service
 To
get a feel for the potential impact of voice technologies, just
look at how popular web-based self-service has become with customers.
Voice technologies now allow companies to provide self-service options
with similar functionality and usability to the web over the telephone.
Well-designed applications not only increase customer choice, in
terms of channels and extended hours, but can also drive down costs
and improve efficiency.
Why All the Interest Now?
Unprecedented numbers of companies are currently investigating or
deploying these technologies. In our view a number of factors are
driving this: Firstly there is pressure to reduce the cost of customer
service whilst maintaining quality, to provide a competitive edge,
or fend off the increasing threat of offshore outsourcing; secondly
the technology can now deliver a really good user experience; finally
the emergence of new open standards for the technology, especially
VXML means that costs have fallen dramatically, and the technology
is accessible to far more users. |
 |
The
Technology Pick And Mix
In order to understand how you can build automated phone-based applications
it is useful to have a basic understanding of the technology components
which are available. An IVR platform is the basic enabler, and all
the manufacturers are now offering systems which are compatible
with the new open standards. The most basic form of interaction
is DTMF (pressing the keys on the telephone handset) – this
is a somewhat maligned approach, but still has a very useful role
to play as a reliable data input method. One can then add speech
recognition, which will allow the system to understand what the
caller is saying.  All
meaningful dialogues are two - way, and the system can respond back
to the caller using pre-recorded prompts, or a Text-to-Speech system,
which enables the “voicing back” of text-strings from
a database. If high levels of security, or fraud prevention are
a requirement, then speaker verification adds an additional safeguard
to ensure the person speaking is exactly who they claim to be. This
can also simplify the identification and verification process. Finally,
if the system is to enable useful transactions to take place, there
will generally be some integration with back-end systems. |
|
The
End of the Line for
Recorded Prompts?
Products recently launched by text-to-speech engine vendors could
signal the end of the line for voice-over artists. It is claimed
that these products allow fine-tuning of text-to-speech output,
such that for predicted phrases, the result is practically indistinguishable
from a human voice. Where companies have a “corporate voice”
it is possible to digitise this, so that it can be used for text-to-speech
output whenever required.
For some companies this approach will be very useful, with the potential
to speed up application deployment and deliver cost savings where
large numbers of prompts, or rapidly changing content is needed.
It may also provide insurance against attacks of laryngitis for
the recording artist. As always one needs to take a balanced view,
and consider the advantages against the extra costs of text-to-speech
licences, and the role that only well-directed voice talent can
play in delivering effective applications and an enjoyable experience
for the caller. |
